The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill uses a dynamic context command in SKILL.md (!node -e ...) that executes at load time to audit the local environment. This script reads the project's package.json for @capacitor/ dependencies and checks for the existence of specific iOS configuration files like SceneDelegate.swift. The command is used solely for project state assessment and does not perform network operations or access sensitive system files.\n- [COMMAND_EXECUTION]: The skill includes instructions to run various shell commands including grep, find, npm, and npx. These are intended for auditing the application's source code for legacy patterns and performing the necessary migrations to the UIScene lifecycle. The scope of execution is limited to the local project directory.\n- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from local project files that are not strictly controlled by the skill itself, such as AppDelegate.swift and Info.plist. Ingestion points: Project source and configuration files. Boundary markers: Not explicitly defined in the audit commands. Capability inventory: Local file system auditing via grep and node, plus capability to update packages and modify project files. Sanitization: Not explicitly present in the data ingestion logic. This creates a surface where malicious instructions embedded in project files could theoretically influence the agent's behavior during the migration process.

capacitor-uiscene-migration